Digital Commons Network^™

Lifelong Machine Learning With Adaptive Resonance Theory, Sasha Petrenko

Doctoral Dissertations

"This publication option dissertation is composed of three papers concerning the study of the problem lifelong machine learning with Adaptive Resonance Theory (ART) algorithms. Lifelong learning (L2) is a challenging machine learning paradigm that both encompasses and formalizes the fields of continual learning and incremental learning. The field is concerned with the mitigation of the phenomenon of catastrophic forgetting whereby learning agents that are faced with incrementally novel information deleteriously overwrite previous knowledge if that learning process is not regularized to counteract this consequence. ART algorithms solve this stability-plasticity dilemma by optimally assigning learning to categories or instantiating new knowledge …

New Topologies For High Voltage Gain Dc-Dc Power Electronic Converters – Theoretical Analysis And Experimental Validation, Saeed Habibi

Doctoral Dissertations

"This dissertation is focused on developing and analyzing new high voltage gain DC-DC power electronic converters. The proposed converters in this dissertation are suitable candidates for connecting low voltage sources such as photovoltaic (PV) panels or fuel cell (FC) stacks to a DC microgrid or DC link of an inverter. The proposed converters provide desired characteristics such as high voltage gain at a low or medium duty cycle values, low voltage stress on power semiconductors, and continuous input current. Initially, the research introduces three novel converter topologies: First, two quadratic converters based on the combination of a three-winding coupled inductor …

Sift Feature-Based Relative Altitude Estimation Enhanced With Siamese Network, Shirin Nasr-Esfahani, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

In GPS-denied environments or when GPS signals are unreliable or unavailable, alternative methods of accurate localization with coordinate generation become critical. To address localization, the scale-invariant feature transform (SIFT) algorithm, along with its numerous adaptations, is extensively utilized in computer vision and remote sensing for matching image features to identify objects and perform localization. This article presents a novel approach for estimating the relative altitude of unmanned aerial vehicles (UAVs) using SIFT features' scale (size), omitting the need for additional data like camera intrinsic parameters, as well as extensive image datasets are also required for training. Furthermore, the approach enhances …

Enhanced Multidimensional Harmonic Retrieval In Mimo Wireless Channel Sounding, Yanming Zhang, Wenchao Xu, A. Long Jin, Tianquan Tang, Min Li, Peifeng Ma, Lijun Jiang, Steven Gao

Electrical and Computer Engineering Faculty Research & Creative Works

This paper introduces a recursive parallel dynamic mode decomposition (RPDMD) scheme tailored for multidimensional harmonic retrieval (MHR), specifically applied to MIMO wireless channel sounding. The RPDMD algorithm is devised to address the complexities inherent in multidimensional scenarios, leveraging the dynamic mode decomposition (DMD) framework within a recursive parallel structure. Initially, the observed tensorial multidimensional harmonic data is transformed into a two-dimensional matrix format along the r-th dimension. Subsequently, DMD dissects this matrix data into eigenvalues and their associated modes. The real and imaginary components of the DMD eigenvalues yield damping factors and frequencies in the r-th dimension, respectively. Furthermore, recursive …

A Compact Frequency Comb Generator For Large-Format Superconducting Microwave Kinetic Inductance Detectors, Sungyoun Seo

Open Access Master's Theses

Microwave Kinetic Inductance Detectors (MKIDs) are promising superconducting detectors that achieve superior sensitivity to silicon-based photo detectors by leveraging a small bandgap of superconductors. They are advantageous for sensitive imaging instruments, particularly space telescopes. However, a large-format array with several hundreds of thousands of detectors is required for large-scale observations in space, and it encounters multiple challenges. The biggest challenge is the absence of compact, energy-efficient, and radiation-tolerant readout solutions. Of multiple functions in the MKID readout, the new frequency comb generator is presented in this thesis. It features a low on-chip memory requirement while promising the relaxed frequency resolution …

Optimal Subspace Estimation For Linear Nested Arrays: Applications And Performance Metrics, Brendan Dunn

Open Access Master's Theses

OSE (Optimal Subspace Estimation) is an algorithm that obtains an estimated subspace from structured data observed in noise. While OSE is able to obtain an accurate subspace estimate with a small number of snapshots, it has not been demonstrated how this affects the performance of many applications that leverage OSE. One such application is beamforming; Using an OSE-based beamformer, this thesis will apply common performance metrics to measure how much of an advantage an accurate subspace estimate provides. This algorithm will be compared against other widely used beamformers such as MPDR and DMR to benchmark its efficacy. Mismatch is then …

An Iterative Shifting Disaggregation Algorithm For Multi-Source, Irregularly Sampled, And Overlapped Time Series, Colin O. Quinn, Ronald H. Brown, George F. Corliss, Richard J. Povinelli

Electrical and Computer Engineering Faculty Research and Publications

Accurate time series forecasting often requires higher temporal resolution than that provided by available data, such as when daily forecasts are needed from monthly data. Existing temporal disaggregation techniques, which typically handle only single, uniformly sampled time series, have limited applicability in real-world, multi-source scenarios. This paper introduces the Iterative Shifting Disaggregation (ISD) algorithm, designed to process and disaggregate time series derived from sensor-sourced low-frequency measurements, transforming multiple, nonuniformly sampled sensor data streams into a single, coherent high-frequency signal. ISD operates in an iterative, two-phase process: a prediction phase that uses multiple linear regression to generate high-frequency series from low-frequency …

Computer Vision-Based Framework For Data Extraction From Heterogeneous Financial Tables: A Comprehensive Approach To Unlocking Financial Insights, Iftakhar Ali Khandokar, Priya Deshpande

Electrical and Computer Engineering Faculty Research and Publications

Information extraction from financial document images is crucial in computer vision and NLP, as financial data often exists in image or PDF format, enabling organizations to analyze and make informed business decisions using OCR advancements. The table contents of financial document images are one of the prominent structures to confine important portions of data of the document and many Deep learning-based methods have been proposed to detect Table regions inside document images. The shortcomings of the current approach are that it is bounded within the detection of the table region and struggles in cases such as handling different layouts and …

Aperiodically Intermittent Dynamic Event-Triggered Control For Predefined-Time Synchronization Of Stochastic Complex Networks, Lei Xue, Haoyu Zhou, Yongbao Wu, Jian Liu, Donald C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

In this paper, the problem of practical predefined-time synchronization in mean square (PTSMS) of stochastic complex networks (SCNs) is investigated through dynamic event-triggered control (E-TC). Different from the existing literature, this paper considers the dynamic E-TC in an a periodically intermittent control framework and employs the average control rate, which makes it easier to satisfy the conditions of the theorem. In comparison to existing finite-time and fixed-time synchronization, by introducing the time-varying function, it can be guaranteed that all states of SCNs achieve the practical PTSMS within a preset time without calculating the convergence time. Combined with stochastic analysis theory, …

A Low-Frequency-Stable Higher-Order Isogeometric Discretization Of The Augmented Electric Field Integral Equation, Maximilian Nolte, Riccardo Torchio, Sebastian Schöps, Jürgen Dölz, Felix Wolf, Albert E. Ruehli

Electrical and Computer Engineering Faculty Research & Creative Works

This contribution investigates the connection between Iso geometric analysis (IGA) and integral equation (IE) methods for full-wave electromagnetic problems up to the low-frequency limit. The proposed spline-based IE method allows for an exact representation of the model geometry described in terms of nonuniform rational B-splines (NURBS) without meshing. This is particularly useful when high accuracy is required or when meshing is cumbersome, for instance, during the optimization of electric components. The augmented electric field IE (EFIE) is adopted, and the deflation method is applied, so the low-frequency breakdown is avoided. The extension to higher-order basis functions is analyzed and the …

Explainable And Safety Aware Deep Reinforcement Learning-Based Control Of Nonlinear Discrete-Time Systems Using Neural Network Gradient Decomposition, Behzad Farzanegan, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents an explainable deep-reinforcement learning (DRL)-based safety-aware optimal adaptive tracking (SOAT) scheme for a class of nonlinear discrete-time (DT) affine systems subject to state inequality constraints. The DRL-based SOAT utilizes a multilayer neural network (MNN)-based actor-critic to estimate the cost function and optimal policy while the MNN update laws are tuned both using the singular value decomposition (SVD) of activation function gradient in order to mitigate the vanishing gradient issue and safety-aware Bellman error at each layer. An approximate safety-aware optimal policy is developed using Karush–Kuhn–Tucker (KKT) conditions by incorporating the higher-order control barrier function (HOCBF) into the …

Adaptive Nussbaum Design For Nonholonomic Systems With Asymptotic Stabilization Against False Data Injection, Guilong Liu, Yongliang Yang, Weinan Gao, Donald C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

This article addresses the stabilization challenges of nonholonomic systems under the threat of false data injection (FDI) attacks, which compromise the integrity of state information. A novel adaptive control strategy using Nussbaum-type gains is proposed to ensure the asymptotic stability of the closed-loop system while maintaining signal boundedness. The approach extends conventional Nussbaum designs to handle multiple unknown control directions. It integrates online learning mechanisms to mitigate the impact of FDI attacks. Additionally, adaptive backstepping and fuzzy-logic systems are utilized to approximate and compensate for unknown nonlinear dynamics. The methodology transforms nonholonomic systems into equivalent cascade structures to address inherent …

Safe Optimal Control Of Quadrotor Formations Using Multilayer Neural Networks And Continual Learning, Ehsan Soleimani, Irfan Ahmad Ganie, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

This article presents an integral reinforcement learning-based optimal formation tracking scheme for multiple quadrotors unmanned aerial vehicles (QUAVs) experiencing nonlinear coupled dynamics and subject to constraints. We use multilayer neural networks (MNN) within an actor-critic framework where the MNN weights are tuned using singular value decomposition (SVD) of the activation function gradient to approximate optimal control policy via backstepping. Additionally, barrier Lyapunov functions (BLF) are introduced to ensure set invariance, thereby maintaining the quadrotors within a defined safety space due to constraints. A novel weight update law for each layer is derived using the HJB approximation error and control input …

Prescribed-Time Nash Equilibrium Seeking For Pursuit-Evasion Game Under Intermittent Control With Undirected/Directed Graph, Lei Xue, Jianfeng Ye, Yongbao Wu, Jian Liu, D. C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

This paper studies the prescribed-time Nash equilibrium (PTNE) seeking problem of the pursuit-evasion game (PEG) with second-order dynamics under the intermittent control (IC) strategy. To achieve Nash equilibrium (NE) in a user-defined prescribed-time, a time-varying high-gain function is incorporated into the design. The core challenge lies in applying IC to NE seeking, which complicates the convergence analysis and control design. To address this sticking point, we construct an auxiliary function and propose a Lyapunov function considering second-order dynamics to solve the PTNE seeking problem of PEG. Building upon the results for undirected graphs, we further extend our findings to directed …

Multi-Model Safe Neuro-Optimal Output Tracking Control Of Autonomous Surface Vessels With Explainable Ai, Behzad Farzanegan, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents a safety-aware deep reinforcement learning (DRL)-based trajectory tracking control of autonomous surface vessels (ASVs). A multilayer neural network (MNN) observer estimates the ASV's state and uncertain dynamics. By utilizing the estimate state vector from the observer, a safety-aware DRL-based optimal policy is formulated using control barrier function (CBF) and Karush-Kuhn-Tucker (KKT) conditions. An actor-critic MNN with singular value decomposition (SVD)-based update mitigates vanishing gradients. To enhance adaptability, an online safe lifelong learning (SLL) scheme counters catastrophic forgetting across varying ASV dynamics. The Shapley Additive Explanations (SHAP) method identifies key features influencing the control policy. Simulations on an …

Enhancing Measurement Accuracy In Industrial Applications: The Impact Of Sensor Data Imputation On Model Parameter Estimation, Steven Thompson, Michkath Omanda Bouraima, Maciej Jan Zawodniok

Electrical and Computer Engineering Faculty Research & Creative Works

Digital twins are meant to revolutionize the manufacturing industry by enabling advanced condition monitoring and predictive maintenance processes. However, disruptions within the manufacturing process, such as sensor malfunctions or connectivity issues are inevitable and will cripple these advanced analysis methods if not properly addressed. Therefore, efficient data management and analysis practices are key to advancing this technology. This work examines the impact of missing data imputation on model parameter estimation, a crucial task in developing models for digital twins. We theoretically derive the Cramer-Rao Lower Bound (CRLB) for a DC signal with an unknown scalar parameter in the presence of …

Reinforcement Learning-Based Nonlinear Optimal Discrete-Time Control Of Power Systems, Vijay Kumar Singh, Behzad Farzanegan, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents a partially model-free adaptive optimal tracking control method for power systems, specifically targeting a synchronous generator connected through a reactive transmission line. By integrating the tracking error dynamics with reference trajectory dynamics, an augmented system is created. A discounted performance function is introduced to address the nonlinear tracking problem optimally. Unlike traditional methods that compute feedforward and feedback terms separately, the proposed approach calculates both simultaneously by minimizing the discounted performance function. The discrete-time tracking Bellman and Hamilton-Jacobi-Bellman (HJB) equations are derived, and a reinforcement learning (RL)-based technique is employed to solve the optimal policy online without …

Distributed Sapphire Fiber Bragg Grating-Based Thermal Profiling Of Submerged Entry Nozzles, Farhan Mumtaz, Hanok W. Tekle, Bohong Zhang, Xiaodong Li, Sunday Abraham, Bryant Mathis, Jeffrey D. Smith, Ronald J. O'Malley, Jie Huang

Electrical and Computer Engineering Faculty Research & Creative Works

This article focuses on the application of sapphire fiber Bragg gratings (FBGs) for instrumentation in submerged entry nozzles (SENs) within the steelmaking industry. The SEN is pivotal for transferring molten steel from a tundish to a mold, while preventing the infiltration of oxygen and nitrogen from the surrounding environment. Maintaining optimal flow conditions in the mold is crucial for ensuring casting process stability and maintaining high-quality steel. Sapphire FBG sensors have been instrumented in SENs to enable distributed thermal mapping for monitoring the health of the SEN. The optical sensor comprises three cascaded sapphire FBGs inscribed using femtosecond (FS) laser …

A Non-Destructive And Simple Setup Method For Dielectric Liquid Characterization In A Wide Frequency Range With Djordjevic-Sarkar Model, Reza Vahdani, Reza Asadi, Seyedmehdi Mousavi, Xiaoning Ye, Donghyun Kim

Electrical and Computer Engineering Faculty Research & Creative Works

This paper introduces a non-destructive and simple setup method for characterizing dielectric liquids over a broad frequency range (up to 30 GHz) using the Djordjevic-Sarkar model. By employing a differential microstrip line and comparing scattering parameters in air-filled and liquid-immersed scenarios, the proposed method achieves precise dielectric constant (DK) and dissipation factor (Df) extraction. 2 Liquid samples (PAO4 and DC-15) were tested using this method. Validation against the cavity resonance method demonstrates a strong agreement for the extracted DK values, with a relative error of less than 1.5 %, indicating high accuracy. However, the method is less sensitive to Df, …

Assessing Techno-Economic Performance Of Synchronous Condensers And Static Synchronous Compensators In Renewable Energy-Integrated Weak-Grids Using Hedge Feedforward Feedback-Based Online Gated Recurrent Unit, Md Ohirul Qays, Iftekhar Ahmad, Daryoush Habibi, Mohammad A.S. Masoum

Research outputs 2022 to 2026

To achieve net-zero targets in many counties, renewable energy generators (REGs) are integrated with grids where less fault current is produced compared to synchronous generators. Accordingly, fault level availability, known as ‘system strength’, is reduced at point of coupling (POC) buses. A minimum system strength level is crucial for REGs to accurately detect and ride-through faults. To maintain adequate system strength in renewable energy-based weak grids, researchers and engineers have recommended supplementary devices, such as synchronous condensers (SynCons) or static synchronous compensators (STATCOMs), to provide the required fault-current. Because SynCons and STATCOMs are costly devices, their optimal sizes and placement …

Vaim-Cff: A Variational Autoencoder Inverse Mapper Solution To Compton Form Factor Extraction From Deeply Virtual Compton Scattering, Manal Almaeen, Tareq Alghamdi, Brandon Kriesten, Douglas Adams, Yaohang Li, Huey-Wen Lin, Simonetta Liuti

Computer Science Faculty Publications

We develop a new methodology for extracting Compton form factors (CFFs) from deeply virtual exclusive reactions such as the unpolarized DVCS cross section using a specialized inverse problem solver, a variational autoencoder inverse mapper (VAIM). The VAIM-CFF framework not only allows us access to a fitted solution set possibly containing multiple solutions in the extraction of all 8 CFFs from a single cross section measurement, but also accesses the lost information contained in the forward mapping from CFFs to cross section. We investigate various assumptions and their effects on the predicted CFFs such as cross section organization, number of extracted …

S²Il: Structurally Stable Incremental Learning, S. Balasubramanian, P. Yedu Krishna, Talasu Sai Sriram, M. Sai Subramaniam, Manepalli Pranav Phanindra Sai, Ravi Mukkamala

Computer Science Faculty Publications

Feature Distillation (FD) strategies are proven to be effective in mitigating Catastrophic Forgetting (CF) seen in Class Incremental Learning (CIL). However, current FD approaches enforce strict alignment of feature magnitudes and directions across incremental steps, limiting the model’s ability to adapt to new knowledge. In this paper, we propose Structurally Stable Incremental Learning (S²IL), a FD method for CIL that mitigates forgetting by focusing on preserving the overall spatial patterns of features which promote flexible (plasticity) yet stable representations that preserve old knowledge (stability). We also demonstrate that our proposed method S²IL achieves strong incremental accuracy and outperforms other FD …

Energy-Based Deep Incomplete Multi-View Clustering, Ziyu Wang, Yiming Du, Rui Ning, Lusi Li

Computer Science Faculty Publications

Incomplete multi-view clustering (IMVC) deals with real-world scenarios where certain views are partially missing, posing significant challenges to effective clustering. Most existing IMVC approaches face a trade-off: imputation-free methods suffer from information bias and imbalance, while full-imputation methods risk introducing and propagating noise. To overcome these limitations, we propose Energy-Based Deep Incomplete Multi-View Clustering (Energy-DIMC), a novel selective-imputation framework that leverages energy-based models (EBMs) to guide reliable imputations and robust clustering. EBMs assess data compatibility by assigning lower energy to more coherent structures, effectively modeling complex inter-view and inter-sample dependencies. Inspired by EBMs, Energy-DIMC integrates four key components: 1) a …

Contextual Memory Recall: A Novel Metric For Class Incremental Learning, Balasubramanian S, Sai Subramaniam M., Sai Sriram Talasu, Yedu Krishna P., Pranav Phanindra Sai M., Darshan Gera, Ravi Mukkamala

Computer Science Faculty Publications

We propose a novel metric for class incremental learning (CIL) called Contextual Memory Recall (CMR), which evaluates how well a CIL model recalls previously learned classes when given relevant past cues. Inspired by human memory, CMR offers newer insights into continual aspects of a CIL model that were not addressed by previously proposed metrics for CIL. Specifically, the standard metric, average incremental accuracy (AIA), overlooks the quality of evolving feature representations, whereas our proposed CMR accounts for it. As a result, methods using feature distillation perform well under AIA but poorly under CMR, while those without feature distillation excel under …

Largely Enhanced Out-Of-Plane Electromechanical Coupling Effects In Two-Dimensional Molybdenum-Disulfide/Boron-Nitride Heterostructures, Qiong Liu, Vijay Kumar Choyal, James E. Morris, Timon Rabczuk, Xiaoning Jiang, Xiaoying Zhuang

Electrical and Computer Engineering Faculty Publications and Presentations

The electromechanical coupling effects in two-dimensional (2D) transition metal dichalcogenides (TMDs) have attracted great interest. However, for 2D TMDs, piezoelectricity is confined to the basal plane, and the flexoelectricity-derived out-of-plane electromechanical response is usually faint, limiting the applications of this material family using the out-of-plane electromechanical effects. Here, this work reports a facile strategy to greatly enhance the out-of-plane electromechanical response of hexagonal molybdenum disulfide (2H-MoS₂) nanoflakes by stacking monolayer hexagonal boron nitride (h-BN) on 2H-MoS₂ nanoflakes to form MoS₂/BN heterostructures. The d^eff/₃₃ coefficient of MoS₂/BN can reach a value comparable …

Estimating Electric Power Consumption Of Grid-Enabled Residential Heat Pump Water Heaters, Dana Paresa, Robert Bass

Electrical and Computer Engineering Faculty Publications and Presentations

We present two methods for estimating compressor electric power consumption of residential heat pump water heaters. Power is calculated using data from voltage and current instrumentation, which add costs and potential failure points to the unit. The two methods presented herein use thermal instrumentation or derived data that are typically included within modern residential heat pump water heaters. One method uses data from lower and upper tank temperature sensors. The other method uses an attribute that is commonly available within grid-enabled flexible load heat pump water heaters. For both methods, fit equations were derived, for which coefficients are tuned to …

Personalized Persuasion In The Digital Age: A Data-Driven Approach To Effective Communication, Annye Braca

Doctoral

This thesis investigates the potential of Machine Learning (ML) to personalize persuasive marketing messages. It explores the identification of individuals receptive to specific persuasion techniques based on their psychometric profiles. By developing ML models that incorporate these profiles, the thesis aims to predict the impact of tailored messages and improve the effectiveness of marketing communication.

Secured Communication And Coexistence For Efficient Spectrum Utilization, Xue Wei

Electronic Theses & Dissertations (2024 - present)

Over the years, the demand for wireless communications has grown exponentially, resulting in a scarcity of available spectrum. This scarcity leads to the need for efficient spectrum utilization, a key factor in ensuring reliable and uninterrupted communications. One of the key challenges in achieving reliable communication is ensuring secure transmissions to protect the privacy and integrity of data for active users. Wireless communications are vulnerable to eavesdropping, interception, and unauthorized access, which can compromise the confidentiality and integrity of transmitted data. The coexistence of multiple wireless communication systems is another challenge to be addressed for passiver users. In many cases, …

Neural Network-Based Image Compression, Atefeh Khoshkhahtinat

Graduate Theses, Dissertations, and Problem Reports (ETD)

The rapid advancement of information technology and the exponential growth of digital communication have significantly increased the demand for efficient data compression techniques that reduce storage requirements, minimize bandwidth consumption, and accelerate data transmission—without substantially compromising data quality. This dissertation addresses these challenges by investigating and developing advanced learned image compression (LIC) methods, with a particular focus on lossy compression for both natural images and scientific imagery obtained from NASA’s Solar Dynamics Observatory (SDO) mission. Traditional image compression standards—such as JPEG, JPEG2000, BPG, and HEVC—rely on manually engineered transforms and heuristic rules, which often lack the adaptability required to accommodate …

Quantum Efficiency, Electron Spin Polarization And Lifetime Study For Gaas Based Truncated Nanocone Array Photocathodes, Md. Aziz Ar Rahman, Greg Blume, Md Abdullah Mamun, Shukui Zhang, Hani Elsayed-Ali

Electrical & Computer Engineering Faculty Publications

[First slide] Direct band-gap photocathode (Eg = 1.42 eV).

Valence band maximum (VBM) is doubly degenerate, Max. electron spin polarization (ESP) ~35-40%.

Strained-supperlattice (SSL) removed the VBM degeneracy, ESP ~85-90% with QE of ~1% at 780 nm.

Requires UHV: Lifetime ~200 C with ~200 μA beam current.

Ion-back bombardment degrades the lifetime.

Improvement in charge lifetime required for JLAB polarized positron source program: 1mA-10mA beam current for 2 weeks ~ 10³10⁴C.